Researchers use different approaches when trying to find the answers to their questions. There are many situations in which researchers do not know exactly what the actual question is (e.g., a discovery scenario), while there are other situations in which even though researchers know exactly what they want to ask (e.g., a known path scenario), they might start at completely different points in order to get to the same result.
Traditional software query environments usually provide a single way to compose the actual question a researcher wants to ask. This is very restrictive, because this is not the natural way researchers tend to think. Further, not all researchers know a specific query language.